Few assumptions about language feel as secure as this one: to understand the meaning of a sentence is to know the conditions under which it would be true or false.
This idea has shaped philosophy, logic, and linguistics for more than a century. It is elegant, powerful, and—within carefully delimited domains—remarkably successful.
It is also deeply misleading.
Meaning is not exhausted by truth conditions.
The Appeal of Truth-Conditional Meaning
Truth-conditional semantics promises clarity. If meaning can be reduced to truth conditions, then understanding language becomes a matter of knowing how sentences map onto states of affairs. Disagreement becomes disagreement about facts. Inference becomes formally tractable. Meaning appears cleanly separable from use.
This picture works well where language has already been constrained to behave propositionally: in mathematics, in logic, in carefully regimented fragments of scientific discourse.
But this success is local.
And its locality is precisely what the picture conceals.
What Truth Conditions Presuppose
For truth conditions to do explanatory work, several things must already be in place.
There must be a stable domain of evaluation. The relevant distinctions must be fixed. The terms must already have settled reference. The situation must be sufficiently determinate that a binary verdict—true or false—makes sense.
These are not achievements of truth-conditional semantics.
They are its prerequisites.
Truth conditions do not generate meaning; they operate within a space of meaning that has already been configured.
Meaning Comes First
Before a sentence can be evaluated for truth, it must first mean something.
That meaning is not propositional in the first instance. It is relational and construal-based. It involves the organisation of experience: what counts as relevant, what distinctions matter, what continuities are assumed, what perspectives are taken up.
Consider again an apparently simple sentence:
The meeting was difficult.
What would its truth conditions be? Difficult for whom? In what sense? Compared to what? Over what stretch of time? These are not missing parameters waiting to be filled in. They are dimensions of construal that the sentence deliberately leaves open.
The sentence means perfectly well without settling them.
Why Non-Propositional Meaning Is Not Defective
Much of ordinary language does not aim at truth evaluation at all.
Questions, commands, promises, apologies, expressions of emotion, assessments, narratives—these do not fail to have meaning because they lack truth conditions. Their meanings operate in different registers: they organise action, coordinate expectation, negotiate stance, and shape shared understanding.
Even declarative sentences often function in ways that outrun truth evaluation. They can soften, provoke, invite, warn, or align without asserting anything that could be cleanly adjudicated as true or false.
To treat such uses as secondary or derivative is to mistake a local formal success for a general theory of meaning.
Truth as a Local Achievement
Truth is not eliminated on this account. It is relocated.
Truth arises when construal has been sufficiently stabilised that evaluation becomes useful. It is an achievement, not a foundation. Within those local conditions, truth-conditional reasoning can be extraordinarily powerful.
But outside those conditions, insisting on truth evaluation distorts what language is doing. It forces meanings into a mould they were never meant to fit.
Truth is something we do with language under certain constraints.
It is not what language is.
After Truth Conditions
Once meaning is understood as prior to truth, a number of familiar confusions dissolve.
Disagreements need not be about facts; they may concern construal. Understanding need not entail agreement. Precision need not entail rigidity. And formal reasoning no longer appears as the ultimate arbiter of meaning, but as one specialised practice among others.
Language is not a device for producing truth-valued propositions.
It is the medium in which spaces of intelligibility are first opened.
In the next post, we will turn to one of the most persistent misunderstandings about natural language—ambiguity—and show why it is not a flaw to be eliminated, but a condition of linguistic possibility itself.
No comments:
Post a Comment